Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Advertise on LowEndTalk.com
What is more processing power - 10x, $5 vps or 1, $50 dedi
New on LowEndTalk? Please read our 'Community Rules' by clicking on it in the right menu!

What is more processing power - 10x, $5 vps or 1, $50 dedi

lowendguy7lowendguy7 Member
edited February 14 in General

I would say the 10x $5 ones would be the clear winner.

Why? Here is my case:

I have an 8core amd 8320 as my home machine, certainly not a top of the line processor but not too shabby. I run that on my encoding script and it will take around 12 hours to complete. This is the same as being a 'dedi' since I am the only one using it.

On the other hand, 1 $5 vps will take around 2-3 days to complete the same script. As such 10 x 2-3 days = way more than just 1 dedi because I doubt anything would beat that for 50 bucks a month. A dedi would need to be at least 10 times faster to be worth buying but for 50 I don't think you could get any server to do that.

Thoughts?

Thanked by 1yoursunny

Comments

  • MikeAMikeA Member, Provider
    edited February 14

    Here is a Geekbench 4 from a full FX 8320 CPU:
    https://browser.geekbench.com/v4/cpu/15219995
    2969 | 11185

    Here is a Geekbench 4 from 4 cores in a VPS from a Ryzen 3950X that cost $20:
    https://browser.geekbench.com/v4/cpu/15208559
    5296 | 16361

    Here is a Geekbench 4 from a 2 core VPS on a Ryzen 3800X that cost about the same:
    https://browser.geekbench.com/v4/cpu/15208509
    5869 | 10352

    So yes, it depends on the VPS hardware.

    Thanked by 1vimalware

    ExtraVM - AMD Ryzen VPS starting @ $3.50
    USA (TX, VA, FL), CA, FR, UK, SGP, AU

  • lowendguy7lowendguy7 Member
    edited February 14

    Thanks, the vps for the $5 ones would be contabo VPS S SSD but do not see an immediate report for that on geekbench on a cursory search. There is one for the M models though.

    @MikeA said:
    Here is a Geekbench 4 from a full FX 8320 CPU:
    https://browser.geekbench.com/v4/cpu/15219995
    2969 | 11185

    Here is a Geekbench 4 from 4 cores in a VPS from a Ryzen 3950X that cost $20:
    https://browser.geekbench.com/v4/cpu/15208559
    5296 | 16361

    Here is a Geekbench 4 from a 2 core VPS on a Ryzen 3800X that cost about the same:
    https://browser.geekbench.com/v4/cpu/15208509
    5869 | 10352

    So yes, it depends on the VPS hardware.

    EDIT: wow, however I see that on this one https://browser.geekbench.com/v5/cpu/918428 which is 6 cores the score is only above 1k! so that means my initial surmise is wrong! and either of the offerings you listed would be a better bet, plus way less hasstle to install just one OS + additional soft vs 10x :hushed:

  • muffinmuffin Member
    edited February 14

    @lowendguy7 said:
    Thanks, the vps for the $5 ones would be contabo VPS S SSD but do not see an immediate report for that on geekbench on a cursory search. There is one for the M models though.

    @MikeA said:
    Here is a Geekbench 4 from a full FX 8320 CPU:
    https://browser.geekbench.com/v4/cpu/15219995
    2969 | 11185

    Here is a Geekbench 4 from 4 cores in a VPS from a Ryzen 3950X that cost $20:
    https://browser.geekbench.com/v4/cpu/15208559
    5296 | 16361

    Here is a Geekbench 4 from a 2 core VPS on a Ryzen 3800X that cost about the same:
    https://browser.geekbench.com/v4/cpu/15208509
    5869 | 10352

    So yes, it depends on the VPS hardware.

    EDIT: wow, however I see that on this one https://browser.geekbench.com/v5/cpu/918428 which is 6 cores the score is only above 1k! so that means my initial surmise is wrong! and either of the offerings you listed would be a better bet, plus way less hasstle to install just one OS + additional soft vs 10x :hushed:

    What you just posted is a Geekbench 5 result, not 4. Numbers on Geekbench 5 are much lower. That looks pretty average meh for a 6 core vps, though, compared to things like ryzen, are slow.

    Thanked by 1MikeA
  • dodheimsgarddodheimsgard Member
    edited February 14

    Dedi is always winner. Been there, done that.

    Whats most importnat, lets say you test some VPS performance, you cant be sure if VPS performance will be same after 2 hours or 2 weeks since resources are shared. You may be on empty node, after 2 weks you may be on full node, results will differ.

    On dedi you know what you get and you will always get same performance.
    Some people use servers with strong CPU for crypto mining. Do you think these guys use VPS? All use dedicated servers. Guess why ;)

    Edit:
    For this price ($50) you could get two or close to two servers with Xeon 1230 or similar CPU that has similar performance to AMD 8320.
    Back in time ive got Dual Xeon 2670v2 for $69 a month. In some CPU heavy tasks (selenium tests) this CPU can beat 3-4 Xeon 1230v2 machines.

  • marvelmarvel Member without signature

    With most providers it's against the ToS anyway to fully load the CPU for days on a $5 VPS.

    Thanked by 1SmallWeb
  • NeoonNeoon Member
    edited February 14

    That's impossible to compare.
    Also, you end up unlikely with dedicated cpu cores on a 5$ VPS if you get dedicated cpu cores they end up crappy less then 1k if even.

    So comparing peak performance to dedicated, makes even less sense.

    Thanked by 1AuroraZ
  • @Neoon said:
    That's impossible to compare.
    Also, you end up unlikely with dedicated cpu cores on a 5$ VPS if you get dedicated cpu cores they end up crappy less then 1k if even.

    So comparing peak performance to dedicated, makes even less sense.

    The whole thread makes no sense at all. Encoding = Dedi period. I am sure who ever he shares a server with would love him pegging resources for hours. Sure there is dedi cores, but not dedi drives. This kind of stuff could bring a whole node to it's knees.

    Catch me over at Primary DNS. If you want to chat I am done with this cesspool.

  • well I been doing this rountine for 2 years or so with many different providers and no complaints and it isn't 24/7 encoding. They will encode then submit the batch and repeat etc so maybe 50/50.> @AuroraZ said:

    @Neoon said:
    That's impossible to compare.
    Also, you end up unlikely with dedicated cpu cores on a 5$ VPS if you get dedicated cpu cores they end up crappy less then 1k if even.

    So comparing peak performance to dedicated, makes even less sense.

    The whole thread makes no sense at all. Encoding = Dedi period. I am sure who ever he shares a server with would love him pegging resources for hours. Sure there is dedi cores, but not dedi drives. This kind of stuff could bring a whole node to it's knees.

  • @lowendguy7 said:
    well I been doing this rountine for 2 years or so with many different providers and no complaints and it isn't 24/7 encoding. They will encode then submit the batch and repeat etc so maybe 50/50.> @AuroraZ said:

    @Neoon said:
    That's impossible to compare.
    Also, you end up unlikely with dedicated cpu cores on a 5$ VPS if you get dedicated cpu cores they end up crappy less then 1k if even.

    So comparing peak performance to dedicated, makes even less sense.

    The whole thread makes no sense at all. Encoding = Dedi period. I am sure who ever he shares a server with would love him pegging resources for hours. Sure there is dedi cores, but not dedi drives. This kind of stuff could bring a whole node to it's knees.

    Depends on the Provider, on the load on the Node .....
    To many variables, which you cannot compare again.

    Of course you can peak on that CPU but its not dedicated.

  • And if these comparisons are 'pointless' then how do I choose the best bang for my buck?

  • @lowendguy7 said:
    And if these comparisons are 'pointless' then how do I choose the best bang for my buck?

    You can compare dedis fine, since they have fixed values you can compare.
    Like the cpu, has a specific benchmark, so you can go out and buy a dedi with the lowest price and the best price/peformance ratio.

    You cannot do that on a VPS.

  • If you are not going to plow the CPU for extended periods, VPS is ok. If you are going to plow the CPU 24/7, dedi is better because your VPS will be suspended.

    Deals and Reviews: LowEndBoxes Review | Avoid dodgy providers with The LEBRE Whitelist | Free hosting (with conditions): Evolution-Host, NanoKVM, FreeMach, ServedEZ | Follow latest deals on Twitter or Telegram

  • If the load is distributable and bursty in nature, then a cluster of VPSes might make sense. For any other CPU-limited operation, dedi is the way to go.

  • jsgjsg Member

    It depends. A dedi usually has more computing power than VPSs plus you have full control. On the other side having everything in one place can be a disadvantage too. For backup for example 2 (or even 3) VPSs are better than one dedi because no matter what redundance a dedi had (e.g. mirrored disks) the upper bounds is the system while multiple VPSs (and of course multiple dedis too) offer even DC redundancy unless one has them all from 1 provider in 1 DC.

    For high and continuous workloads most VPSs are not even an option. On the other hand putting all ones name servers on a single dedi, no matter whether in VMs or not, would be stupid.

    Thanks no.

  • @jsg Yeah, there are very few circumstances under which a VPS is better than a dedi, but it's not impossible. Some shitty single thread software that can be run in a cluster has the potential to run better over a few 1-2 vCore VPSes rather than one big 8+ thread dedi.

  • jsgjsg Member

    @AC_Fan said:
    @jsg Yeah, there are very few circumstances under which a VPS is better than a dedi, but it's not impossible. Some shitty single thread software that can be run in a cluster has the potential to run better over a few 1-2 vCore VPSes rather than one big 8+ thread dedi.

    Rarely, but yes. Simple reason: everything on a node is shared, including the caches, storage, etc. Plus virtualization itself eats resources too. So, unless the dedi is paltry and the VPS is a full modern coreor two on say a Zen2 the VPS will usually not be faster. Plus even a full core based VPS will usually not allow you 100% usage.

    But frankly IMO the true and most sensible answer to your OP question (encoding) is "neither. Use your desktop".

    Thanks no.

  • In most cases a $5 Ryzen VPS will outperform a $20 dedi..

    Purveyor of high quality potassium

  • @dahartigan said:
    In most cases a $5 Ryzen VPS will outperform a $20 dedi..

    What evidence you have to support that?

Sign In or Register to comment.